ANTENNAS AND PROPAGATION FOR BODY CENTRIC
COMMUNICATIONS
Hall P S 1 and Hao, Y 2
1
University of Birmingham, Edgbaston, Birmingham, UK, p.s.hall@bham.ac.uk
2
Queen Mary, University of London, Mile End Road, London, UK, y,hao@elec.qmul.ac.uk
ABSTRACT
Body centric wireless communication is now
accepted as an important part of 4th generation
(and beyond) mobile communications systems,
taking the form of human to human
networking incorporating wearable sensors and
communications. There are also a number of
body centric communication systems for
specialized occupations, such as paramedics
and fire-fighters, military personnel and
medical sensing and support.
To support these developments there is
considerable ongoing research into antennas
and propagation for body centric
communications systems, and this paper will
summarise some of it, including the
characterisation of the channel on the body, the
optimisation of antennas for these channels,
and communications to medical implants
where advanced antenna design and
characterisation and modelling of the internal
body channel are important research needs. In
all of these areas both measurement and
simulation pose very different and challenging
issues to be faced by the researcher.
1 INTRODUCTION
Body centric wireless communication is now
accepted as an important part of 4th generation
(and beyond) mobile communications systems
and will be part of the forthcoming
convergence and personalization across the
various domains, which include personal area
networks, (PANs), and body area networks,
(BANs), as noted in a forthcoming book, [1].
Advancements in the miniaturisation of
wearable hardware, embedded software, digital
signal processing, and biomedical engineering
have made possible human to human
networking incorporating wearable sensors and
communications. This can be seen as a
continuation of a trend spearheaded by the
mobile phone, which, over the last few
decades has become smaller and more
convenient for personalized operation.
Alongside this trend, there have been a number
of body centric communication systems for
specialized occupations such as paramedics
and fire-fighters, as well as military personnel.
Use for medical sensing and support, with
either skin mounted sensors or implants, is also
attracting much attention.
A parallel research activity is that of wearable
computers, where the emphasis is on high
levels of computational power, coupled with
sensors and interface equipment. Currently
these are bulky and have wired connections.
The use of wireless connections in these
systems is desirable although there is also a
trend to the development of wired garments,
and the fabric antennas noted below are part of
this.
Body centric communications takes its place
firmly within the sphere of personal area
networks and body area networks (PANs and
BANs). The content of a PAN or BAN
contains a range of communications
requirements. These can be classified as
communications from off-body to an
on-body device or system off-body
communications within on-body
networks and wearable systems on-body
communications to medical implants
and sensor networks in-body
All of the antenna and propagation studies for
personal mobile communications come within
the fist class. What is now of great interest is
the second and third. Of course, in an
integrated system, all may communicate with
each other and the boundaries will become
blurred. Thus, whilst this is not a perfect
subdivision, it does serve to highlight some of
the different challenges for antennas and
propagation in the body centric system.
2 ON BODY SYSTEMS
2.1 Radiowave Propagation
At low frequencies, electromagnetic energy
has a significant penetration depth, and the
body can be used to support communications
channels, [2-4]. For example at 10MHz the
penetration depth is about 200mm for muscle
and over 1 m for fat. At 2.45GHz the depths
are 25 and 120mm respectively. The range of
communication is in practice constrained to
touch range and hence limits its usefulness.
PAN systems, in which user touch or
handshake, is used for low data rate
communication and has been demonstrated, [4],
at 10MHz. At higher frequencies, propagation
has been characterised at ISM and other bands,
for example, 400, 900 and 2450MHz, [5], to
produce simple analytical models for
stationary bodies. The measurements were
taken with dipoles oriented vertically to the
body surface, and were taken over an angle of
180 around the body. The propagation mode
is assumed to be a creeping wave which
exhibits an exponential decay of power, [6].
The expressions are of two slope form with a
breakpoint part way around the body. Beyond
the breakpoint angle, the two rays are assumed
to interfere, resulting in smaller decay but
increased variability. The decay coefficient at
2.4 GHz corresponds approximately to 2
dB/cm, assuming that the body radius is 20 cm.
Significant interference takes place beyond the
breakpoint at 400 and 900 MHz, indicating the
presence of multipath effects, although
multipath is seen to be much smaller at 2.4
GHz.
Measurements of channel loss, for many body
postures and antenna positions, have recently
been made at 2.45GHz, [7, 8]. Fig 1 shows the
results collated to allow separation into classes
of on-body propagation path gain - distance
variation. Two types of path gain variation
with the distance can be roughly distinguished
based on the propagation scenario, namely,
variation in the illuminated region and in the
shadow region. The data points for the
propagation scenarios in which a line of sight
(LOS) was present, (receiver in the illuminated
region), are shown in Fig. 4 by + signs. The
path gain of the antennas in free space in the
ground plane direction relative to each other,
given by the Friis formula, is also shown for
comparison as a solid line. Fig. 4 shows that
most of the LOS data points follow the free
space curve but are generally below it with the
mean difference between the measured and
theoretical values of 5.1 dB. Thus, the path
gain in the illuminated region at 2.45 GHz is
given by the formula:
G P [dB] 5.33 20 log 10 d [cm] (1)
The standard deviation of the difference
between the measured LOS values and those
given by this formula is 4.2 dB. This deviation
can be attributed to the mutual orientation of
the antennas, which are not normally co-
polarized and in the boresight direction of each
other. Occasionally this leads to very large (up
to 25 dB) discrepancies between the data and
the free space curve. The complexity of
determining the relative orientation of the
antennas along the on-body propagation path
renders it impractical to attempt a more
accurate theoretical prediction model, as the
radiation pattern contribution and polarization
mismatch could not be measured sufficiently
accurately.
The data points represented by circles in Fig. 4
correspond to the scenarios when there was no
LOS between the antennas, and the dominant
propagation path corresponded to a creeping
wave propagating around the body. Generally
speaking, the distance variation of such a wave
is fairly complex and depends on a number of
parameters. For example, the distance variation
is different on the free space sections of the
propagation path and on the sections
conformal to the body. However, because the
human body has a complex geometry, which is
highly variable between individuals, tracing
propagation paths around it is difficult and
often inaccurate. Nevertheless, the non-LOS
data points do appear to follow a trend shown
as a straight line in Fig. 4. This trend
corresponds to an exponential attenuation
according to the following linear regression
formula:
G P [dB] 0.36d [cm] 35 (2)
where d is the antenna separation distance
represented in centimeters and G P is the path
gain in dB. The actual data points are spread
around this line with a standard deviation of
5.6 dB.
A number of propagation scenarios could not
be clearly identified as either LOS or non-LOS
because the receiving antenna was very close
to the shadow boundary. Therefore, with a
slight change of body posture it could become
either obstructed (completely or partially) or
unobstructed by a body part such as the trunk
or an arm. These transition data points are
shown in Fig. 4 as x marks. Some of them
seem to follow the same trend as the LOS data
while others are clustered together with the
non-LOS data points.
 -10
-20
-30
Path Gain (dB)
-40
PG[dB] = - 0.36*d[cm] - 35
-50
-60
-70
-80
0 10 20
Fig. 1. Classes of on-body propagation path gain distance variation
(Frequency =2.45GHz, path gain measured with quarter wavelength monopoles)
Recent results at the University of Birmingham
show that there is a significant multipath on
the body at 2.45 GHz, in an anechoic
environment. Fig 2 shows the simulated signal
received by an antenna on the back from an
antenna on the front, transmitting a wideband
pulse. The arrival of signals at three different
times can be seen. An associated experiment,
in which an array was rotated in front of the
body about an axis normal to the body surface
showed that strong signals were arriving from
a number of different directions, including the
shoulders, the waist and from between the
legs. Measurement of the physical length of
these paths and subsequent estimation of the
path delays showed encouraging agreement
with the timings noted in Figure 2. A diversity
Fig 2 Signal received at antenna on back of body
from antenna on front transmitting wideband
pulse (Simulation with XFDTD using Norman
voxel phantom)
Free Space (theory)
LOS Data
Transition Data
Non-LOS Data
Non-LOS Linear Fit
30 40 50 60 70 80 90 100 110
Distance (cm)
gain of up to 6 dB has been achieved with a
pattern diversity antenna at 2.45GHz, [9].
Measurement of the diversity gain of pairs of
monopole antennas as diversity receivers, is
now under way in collaboration with the
University of Pisa. Initial results show that
diversity significant gain can achieved, of the
order of that noted above.
Ultra wideband signals will have some
significance in on body communications and
there is much work underway to characterize
the path loss and dispersion properties, [10,11,
12]. On-body propagation channel measured
was performed using a cpw-fed UWB antenna
[10]. For this measurement, a matrix of
measurement points is used for reliable and
efficient channel characterisation and
modelling with minimum distance between Tx
and Rx of 10cm. Different antenna positions
and various angular orientations are applied
with reference to the transmit antenna on the
right waist. When comparing impulse
performance of the antenna placed on the body
at different angular position, a minimum
fidelity of 60% can be obtained. The
probability density function of measured path
loss data for all body postures and all antenna
orientations fitted well to a normal Gaussian
distribution with a high scaling ( =14.5) value.
This variation is due to body geometry
changes, including angular positions, which
verifies the rapidly changing on-body
environment. To provide on-body channel
models applicable to system designs from

Fig 3. Measured and modelled on-body path loss
with CPW-fed UWB antenna, multi-slope fitting
large-scale analysis, the measured path loss
data is fitted using linear power law. However,
to efficiently model the path loss a multi-slope
approach is required with specified breaking
points at which the received power decays is
different speeds. Figures 3 and 4 show the
measured and modelled multi-slope on-body
path loss and excess time delay respectively.
The high value of the path loss exponent in the
first region is due to the smaller distances
between Tx and Rx. Then the communication
in the following region is assumed to be far
field. Finally, at distances where the Rx is on
1
Mean excess delay
0.9 Normal fit
0.8
0.7
0.6
CDF
0.5
0.4
0.3
0.2
0.1
0
1 2 3 4 5 6 7 8
Mean excess delay (nsec)
Fig 4. Measured and modelled on-body mean
excess delay with CPW-fed UWB antenna
(Threshold used for detecting mutlipath
components is -25dB below the strongest path
level for time window of around 30 nanoseconds)
the side and the back of the body, it is mainly
non-line of sight link and thus surface or
creeping waves, caused by the body
shadowing.
2.2 Wearable Antennas
One of the dominant research topics in
antennas for body centric communications is
wearable, fabric based antennas. Early interest
lead to the development of an RF helmet
antenna [13,14] and an RF vest antenna
[15,16], operating over a wide frequency band.
Since then, the research on wearable antennas
has received significant research interest,
[17,18]. The wearable antenna has come to be
defined as an antenna which is designed and
meant to be a part of clothing [19], although it
is clear that antennas will be worn on the body
in equipment placed in pockets or attached to
the body directly, such as Bluetooth headsets
or medical sensors discussed in section 3. Of
course wearable antennas may be used for
links to antennas off the body, such as mobile
communications base stations, or satellites, for
links to other wearable antennas on the body,
or communicating to medical implants.
There are several, significant challenges in the
design of wearable, fabric based antennas, that
relate mainly to the use of fabrics. They are
prone to bending, flexing and wrinkling, and
the antennas must remain operational under
these circumstances. Fabrics generally have a
low dielectric constant which helps in getting
wide bandwidth from the antenna. Either
conducting films or conducting fibres may be
used. In the latter case, fibre conductivity and
the weave patterns may be important. Fig 5
shows a wearable patch antenna constructed
from copper foil and fleece fabric, [20]. The
fleece, serving as the dielectric in this patch
antenna was 3.5mm thick and the ground plane
was 110 x 130 mm. For this thickness of
material the bandwidths were slightly lower
than the GSM 1900 and WLAN allocations,
indicating the need for thicker fleece material
9 10
Fig 5 Dual band fleece fabric wearable
antenna for 1.9 and 2.4GHz
The performance of various conventional, non
fabric, antennas, including monopole, patch
and PIFA, operating in on body
communications links, has been examined. In
such paths it is important to have a pattern and
polarisation that support propagation close to
the body. However, in any given situation, the
path loss is dependent on the antenna type and
the posture of the body, which determines the
antennas spacings and mutual orientation. In
most positions on the body and for most
postures, the monopole, oriented with the
ground plane parallel to the body surface,
gives the least path loss. However choice of
antenna will also depend on the size available
in the equipment being used for the application
in mind.
3 TELEMEDICINE SYSTEMS
Antennas and propagation for telemedicine
systems can be considered in two parts, those
for systems outside the body and those that
communicate with internal implanted sensors
and devices.
A good introduction and overview of wireless
telemedicine can be found in [21]. While the
antennas and propagation aspects of
telemedicine technology have similarities with
other body-centric applications, transmitted
power levels are generally much lower and
antenna efficiencies reduced because of battery
and packaging constraints. Figure 6 illustrates
the concept of a fully-connected patient-centric
wearable telemedicine system, highlighting
both existing and emerging technologies. In
this approach a network of low power
wearable devices (sensors or actuators, surface
worn or implanted) is established using short
range UHF radio. However, these devices may
not have sufficient data processing, energy or
memory resources to fully realise their
function and so a more substantial controller
Data controller /
storage with WLAN /
GPRS etc. backhaul
wearable sensor
ISM / MEDS
medical implant
MICS / MEDS
body or patient
area network personal area
network
Fig 6 Wearable telemedicine system
WLAN wireless local area network
GPRS General Packet Radio Service
ISM Industrial, Scientific and Medical
MEDS Medical Data Service
MICS Medical Implant Communications
Service
0 dBi
+10 dBi
Figure 7 Azimuthal radiation pattern for hip-
worn 5.2 GHz telemedicine unit.
or basestation device may also be part of the
wearable network.
In terms of antennas at frequencies below 1
GHz a small loop is a reasonable choice for
compact body mounted telemedicine
equipment. At higher frequencies monopoles,
PIFAs and dipoles have been used, and in
these cases the body will have a significant
effect. Figure 7 shows an FDTD simulation of
the azimuthal radiation pattern of a hip worn
telemedicine unit operating at 5.2 GHz, [22].
The model included an anatomically realistic
human body phantom, a conducting box
(representing the patient unit) and a thin-wire
dipole antenna. The overall FDTD grid was
499 x 93 x 154 with cubic 3.6 mm voxels. The
body phantom was for a 1.75 m tall adult male
and incorporated 21 tissue types. The sleeve-
dipole antenna used in measurements was
modelled as a centre-fed 25.2 mm (0.36 mm
radius) thin-wire element and was positioned
with a minimum antenna-body spacing of 14.4
mm. The high degree of separation (> 2)
reduced the overall body losses, with a
corresponding FDTD-computed radiation
efficiency of 83.3 % at 5.2 GHz. The
computed pattern was strongly directional,
with a peak gain of +6.0 dBi and a through-
body null of 37.9 dBi in the azimuthal plane
.
Pacemakers and implantable RFID use the
inductive link with a carrier frequencies
between 9kHz to 315KHz with a data rate of
up to 512kb/s. Use of the Medical Implant
Communication System (MICS), [23], at 402
to 405 MHz, allows bands of 300kHz to be
achieved. Examination of the electric and
magnetic components of the field, when a
plane wave meets the body, show that there is
a node of the electric field and the anti-node of
the magnetic field at the surface of the body.
Inside the body there is a dominating
propagating wave which is attenuated due to
the conductivity of the muscle tissue. The
magnetic field is strong at the surface the body
which implies that a magnetic antenna would
be beneficial for subcutaneous implants. RF
telemetry reception from implanted sensors has
been demonstrated by several researchers [24,
25].
For deeper implants various antenna types
have been examined, such as the long wire
antenna, [26], the patch or PIFA, [27] and the
coil [25], although it has been shown that in
the body environment a magnetic based
antenna will still have significantly better
efficiency than an electric one, and that
increasing the insulation diameter around the
antenna will also increase its efficiency.
4 THE FUTURE
For on- or in-body environments, antennas are
required to be conformal and immune from
frequency and polarisation detuning. It is also
likely that the associated electronics and
sensors will be conformal in shape. Whilst
many antennas in wearable equipment will be
of conventional miniaturised form, some will
be made using textiles, particularly for
specialist occupations such as firefighters and
paramedics, and also where fashion dictates
form and design. Antennas for medical
implants are particularly challenging due to a
very variable environment and the need for
low frequency operation using small antennas.
Conventionally, in antenna design, physical
dimensions, directivity, efficiency and so on,
are constrained by the wavelength of the
radiation involved. These constraints are very
relevant to body centric antenna applications.
The use of medical implants will increase as
the number and capability of nano and micro
sensors and devices increases.
Communications between implants may then
be used to form networks, which improve the
system functionality. It could be said that
system design will be unique to the implant
function required, but for future design and
production efficiency through commonality,
more systematic studies of antennas and
propagation for implant are required.
Although there has been significant progress in
simulation tools, there is still a need for
improvement particularly in run time for full
body simulators.
REFERENCES
1. Hall, P S and Hao, Y, (eds),
Antennas and Propagation for Body Centric
Wireless Communications, (Artech House,
Boston/London, 2006)
2. T. G. Zimmerman, "Personal Area
Networks: Near-Field Intra-Body
Communication," IBM Systems Journal,
vol.35, no.3&4, pp.609-617, 1996.
3. T. G. Zimmerman, J. R. Smith, J. A.
Paradiso, D. Allport and N. Gershenfeld,
Applying Electric Field Sensing to Human-
Computer Interfaces, Proc. of CHI95,
pp.280-287, May 1995.
4. K. Fujii, K. Ito and S. Tajima, "A
study on the receiving signal level in relation
with the location of electrodes for wearable
devices using human body as a transmission
channel," Proc 2003 IEEE Int Symp on Ants
and Prop pp.1071-1074, Columbus, OH, USA,
June 2003.
5 Ryckaert, J, De Doncker, P, Meys, R,
de Le Hoye, A and Donnay, S, Channel
model for wireless communication around
human body, Electronics Letts, 29th April
2004, Vol. 40, No. 9, pp 543-544
6 McNamara, D.A., et al.: Introduction
to the uniform geometrical theory of
diffraction (Artech House, Boston, 1990)
7 Nechayev, Y I, Hall, P S,
Constantinou, C C, Hao, Y, Alomainy, A,
Dubrovka, R and Parini, C, On-body path
gain variations with changing body posture
and antenna position, IEEE AP-S
International Symposium, Washington, USA,
July 2005, pp731-734
8 Hao, Y, Alomainy, A, Zhao, Y and
Parini, C, Nechayev, Y I, Hall, P S,
Constantinou, C C Antennas and Propagation
for Body Centric Wireless Communications,
International Wireless Summit 05, Aalborg,
Denmark, September 2005
9 Kamarudin, M R and Hall, P S,
Diversity measurement of antennas in the on-
body environment, EuCAP 2006, Nice, 6-10
November 2006-07-18
10 Alomainy, A.; Hao, Y.; Parini, C.G.;
Hall, P.S.; Comparison between two different
antennas for UWB on-body propagation
measurements, Antennas and Wireless
Propagation Letters Volume 4, 2005
Page(s):31 34.
11 Alomainy, Y. Hao, X. Hu, C. G.
Parini and P. S. Hall, UWB On-Body Radio
Propagation and System Modelling for
Wireless Body-Centric Networks, IEE
proceedings on communications, special issue
on UWB, 2006.
12 Yan Zhao, Yang Hao, Akram
Alomainy, and Clive Parini, UWB On-Body
 Radio Channel Modelling Using Ray Theory requirements, and test methods., European
and Sub-band FDTD Method, Microwave Telecommunications Standards Institute, 2002.
Theory and Techniques, IEEE Transactions on, 24 J.A. Von Arx and K. Najafi, On-
Volume 54, Issue 4, Part 2, April 2006 Chip Coils with Integrated Cores for Remote
Page(s):1827 1835 Inductive Powering of Integrated
13 J. Lebaric, A.-T. Tan, Ultra- Microsystems, 1997 Inter. Conf. Solid-
Wideband RF Helmet Antenna, IEEE State Sensors and Actuators,
MILCOM 2000, Vol. 2, 2000, pp. 591 594. TRANSDUCERS 97, Chicago, IL,
14 J. Lebaric, A.-T. Tan, Ultra- pp. 9991002, June 1619, 1997.
Wideband Conformal Helmet Antenna, IEEE 25 R.N. Simons, D.G. Hall, and F.A.
Asia Pacific Microwave Conference, 2000, pp. Miranda, RF Telemetry Systems for an
1477 1481. Implantable Bio-MEMS Sensor, 2004 IEEE
15 R. Abramo, R. Adams, F. V. Canez, MTT-S Inter. Microwave Symposium Dig.,
H. Price, P. Haglind, Fabrication and Testing Vol. 3, pp. 1433-1436, Fort Worth, Texas,
of the COMVIN Vest Antenna, IEEE June 6-11, 2004.
MILCOM 2000, Vol. 2, 2000, pp. 595 598. 26 C. A. Caceres, Ed., Biomedical
16 J. E. Lebaric, R. W. Adler, M. E. Telemetry. Academic Press, 1965.
Limbert, Ultra-Wideband, Zero Visual 27 J. Kim and Y. Rhamat-Samii,
Signature RF Vest Antenna for Man-Portable Implanted antennas inside a human body:
Radios, IEEE MILCOM 2001, Vol. 2, 2001, Simulations, designs and characterizations,
pp. 1291 1294. IEEE Transactions on Microwave Theory and
17 Salonen, P., Hurme, H., A Novel Technoques, vol. 52, no. 8, pp. 19341943,
Fabric WLAN Antenna for Wearable Aug. 2004.
Applications, 2003 IEEE Antennas and
Propagation Society International Symposium,
Columbus, Ohio, USA, vol. 2, 2003, pp. 700
703.
18 M. Tanaka, J.-H. Jang, Wearable
Microstrip Antenna, 2003 IEEE Antennas and
Propagation Society International Symposium,
Columbus, Ohio, USA, vol. 2, 2003, pp. 704
707.
19 L. Sydnheimo, M. Keskilammi, M.
Kivikoski, A Small Planar Inverted-F
Antenna for Wearable Applications, The
Third International Symposium on Wearable
Computers, 1999, pp. 95 100.
20 P. Salonen, Rahmat-Sami, Y, Dual
band wearable antenna, 200 IEEE Antennas
and Propagation Society International
Symposium, Monterey, California, USA, June
20-25 2004.
21 C.S. Pattichis, E. Kyriacou, S.
Voskarides, M.S. Pattichis, R. Istepanian and
C.N. Schizas, Wireless telemedicine systems:
an overview, IEEE Antennas & Propagation
Mag., Vol. 44, 2, pp. 143153, April 2002.
22 K.I. Ziri-Castro, W.G. Scanlon and
N.E. Evans, Indoor radio channel
characterisation and modelling for a 5.2-GHz
bodyworn receiver, IEEE Antennas and
Wireless Propagation Letters, Vol. 3, 11, pp.
219222, 2004.
23 ETSI EN 301 839-1 Electromagnetic
compatibility and Radio spectrum Matters
(ERM);Radio equipment in the frequency
range 402 MHz to 405 MHz for Ultra Low
Power Active Medical Implants and
Accessories;Part 1: Technical characteristics,
including electromagnetic compatibility

All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.